Spatial Reasoning On 6 Dof Spatialbench
Metrics
Orientation-abs
Orientation-rel
Position-abs
Position-rel
Total
Results
Performance results of various models on this benchmark
Model Name | Orientation-abs | Orientation-rel | Position-abs | Position-rel | Total | Paper Title | Repository |
|---|---|---|---|---|---|---|---|
| SoFar | 31.3 | 54.6 | 33.8 | 59.6 | 43.9 | SoFar: Language-Grounded Orientation Bridges Spatial Reasoning and Object Manipulation | |
| SpaceMantis | 25.0 | 27.2 | 29.2 | 33.6 | 28.9 | SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities | - |
| SpaceLLaVA | 24.9 | 30.9 | 30.5 | 32.4 | 28.2 | SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities | - |
| RoboPoint | 25.8 | 33.8 | 30.8 | 43.8 | 33.5 | RoboPoint: A Vision-Language Model for Spatial Affordance Prediction for Robotics | - |
| SpatialBot | 22.9 | 39.6 | 21.6 | 50.9 | 32.7 | SpatialBot: Precise Spatial Understanding with Vision Language Models | |
| LLaVA-1.5 | 25.8 | 28.3 | 24.5 | 30.9 | 27.2 | Improved Baselines with Visual Instruction Tuning | |
| GPT-4o | 25.8 | 44.2 | 28.4 | 49.4 | 36.2 | GPT-4o System Card | - |
0 of 7 row(s) selected.